Enhanced Static Mixer Design Analysis in Lattice Boltzmann Solver

Download or read book Enhanced Static Mixer Design Analysis in Lattice Boltzmann Solver written by Robert James Strong and published by . This book was released on 2020 with total page 96 pages. Available in PDF, EPUB and Kindle. Book excerpt: The mixing industry has long depended on scaled down experimental methods combined with computational analysis to determine rotating mixer designs for customer applications. Most industrial mixing companies have the capabilities in-house to perform these experiments and the analysis to show customers the benefits of proposed designs. Experimental methods center around the calculation of power draw of the mixing unit, determined from a simple torque cell to determine power draw, and the blend time, shown through acid-base neutralization, which are both fairly simple to calculate from an scaled down rig and apply it to either customer designs or in the development of new mixers. The computational analysis centers around research done by mixing forefathers who developed methodology to calculate time dependent mixing parameters, like blend time, through steady state analysis due to restrictions in computational capacity. This is possible because the majority of the mixing can be observed from studying the macro-scale interactions. The impellers and baffles in a tank drive large scale motion which blends two different species or temperatures together to create a uniform mixture. Similar to rotating mixers, there are two main parameters used when analyzing static mixers. Similar to power draw for rotating mixers, static mixers are driven by the pressure drop across the mixer. The second parameter that is used to determine the effectiveness of a static mixer is the coefficient of variation, a statistical measurement of the degree of uniformity. This looks at a two-dimensional plane located downstream from the mixer outlet and determines the effectiveness of the mixer. This has been used for decades to provide the target for customer designs, but it provides a limited picture of the process. The snap shot on a two-dimensional plane provides a small window into what is happening in the entire process. The coefficient of variation also is merely a statistical parameter that does not provide any actual physical information about the mixing process. Analysis has been performed that has demonstrated the need for more than CoV alone to fully describe the mixing process [26].The progress and continued development of rotating mixers and application analysis is quite different from that of static mixers. Static mixers have a more difficult analysis path than rotating mixers, which has lead to stunted development and growth by the industrial leaders due to number of factors. The first hurdle that mixing companies face is the physics driving the mixing itself. Static mixers drive mixing on the micro-scale. This means that counter to the large sweeping mixing that occurs in stirred vessels, static mixers drive mixing by introducing turbulence into a system which cause two distinct species or temperatures or conditions to interact with one another. The micro-scale mixing provides a hurdle for experimental setups because static mixer test rigs tend to be very large. Mixer performance is dependent upon the length and size of the mixer, which can lead to mixer chains that are the size of a room or larger. The second issue with static mixers is the equipment needed to determine the mixing is expensive and complex. This causes industrial companies to contract out third parties, either consultants like BHR Group or academic institutions to perform analysis. The issue that has arisen with third party experimental testing, is there are not any standards for data acquisition. The difficulty to analyze static mixers experimentally would tend to lead naturally to the exploration of computational analysis of static mixers. Unlike rotating mixers, the microscale mixing of static mixers requires more refined transient computational analysis than rotating mixers, which is difficult with traditional computational finite volume approaches. Traditional solvers require large high performance computers to perform the analysis, which is both expensive and time consuming from a use standpoint. This is why most development has occurred by either private consulting firms like BHR Group, or universities who have large endowments and student researchers who are able to spend the necessary time on computationally expensive analysis. The computational time required to reach a solution is highly prohibitive in an industrial setting. Most results are needed in a relatively short timeframe, and traditional methods require too much time and are too expensive to increase efficiency to be used practically for customer applications. This has led to a limited amount of progress being made from a development standpoint and process design rules based on experience to determine appropriate mixer design for customer applications. This has stunted the progress of static mixer development and innovation as well as reduced the likelihood of implementation in a customer's process when compared to rotating mixers. The past few years have seen a commercially available non-traditional solver, MStar CFD, be developed specifically for use in the mixing industry. This solver was developed with the help of the leading industrial mixing companies, and shows promise in terms of both accuracy, as well as computational speed. The new solver is a lattice Boltzmann based solver, rather than a traditional finite volume solver like ANSYS Fluent, which is traditionally the solver of choice in the mixing industry. The emergence of newly available technology should cause companies to reevaluate their current procedures and open the door for improving the products and processes provided to customers. This paper compares the new non-traditional solver to one of the more popular traditional solvers for accuracy and computational speed in out-of-the-box conditions. The study was designed to test which solver was better equipped for modeling practical mixing cases, and which should be the standard going forward. MStar CFD was found to produce results faster and more accurately when analyzing common mixing scenarios which should cause more companies to switch the industry standard CFD software. Once the solver comparison was completed, it was then deemed important to implement the parameters described by Kresta et al. to improve current design practices that strictly use the coefficient of variation for determining how a system is behaving. The other two parameters, clustering and exposure, are verified and validated within the MStar CFD framework and can be used with the coefficient of variation to more fully describe the mixing in a system. This is important for many different industries as well as for static mixer development because each process may be driven more heavily by a different aspect of mixing. For example, the chemical process industry might need a consistent residence time in order to minimize the production of secondary and tertiary products. In processes where two products are being mixed with very different viscosities, then the clustering will be more important than coefficient of variation in describing what is happening in the process. The third parameter, exposure, can be used to determine the mass transfer rate between the different species in the system. This is important because it can aid in reaction kinetic calculations and predicting total product output from a given setup. These parameters were successfully demonstrated within the MStar CFD framework and should begin to be used immediately within computational mixing circles.